Blind equalization/decoding using low-density parity check codes

In 1948, Claude E. Shannon with his Channel Coding Theorem predicted the existence of long and random codes that can achieve communication with arbitrarily small error probability and finite rate of transmission. The search for codes predicated by Shannon has resulted in several classes of error control codes over the years. There are very few techniques in coding literature that can decode very long codewords with reasonable complexity. With the advent of Turbo Codes in 1993, it has been possible to realize algorithms for decoding very long codewords. Since then, the Turbo idea has been extended to several aspects of the receiver.; In this thesis, we realize a Turbo equalization scheme that can perform blind equalization of a time-varying channel. Adequate mathematical foundations are laid in order to achieve such a system. The receiver presented here consists of three parts: (1) an equalizer that uses a priori information, (2) an expectation-maximization-based blind channel estimator, and (3) a low density parity check (LDPC) decoder that provides error control and the probabilities used in channel estimation and equalization. The novel contribution of this thesis is the generation of joint probabilities from the decoder that is used in the channel estimation, thus completing the missing link in the Turbo equalization.